Add support for linguist-detectable and linguist-documentation (#29267)

KN4CK3R · web-flow · commit 2a278b996fd6 · 2024-02-23T17:24:27.000Z
Add support for `linguist-detectable` and `linguist-documentation` Add tests for the attributes https://github.com/github-linguist/linguist/blob/master/docs/overrides.md#detectable https://github.com/github-linguist/linguist/blob/master/docs/overrides.md#documentation
diff --git a/modules/git/repo_attribute.go b/modules/git/repo_attribute.go
@@ -11,6 +11,7 @@ import (
 	"os"
 
 	"code.gitea.io/gitea/modules/log"
+	"code.gitea.io/gitea/modules/optional"
 )
 
 // CheckAttributeOpts represents the possible options to CheckAttribute
@@ -291,7 +292,7 @@ func (repo *Repository) CheckAttributeReader(commitID string) (*CheckAttributeRe
 	}
 
 	checker := &CheckAttributeReader{
-		Attributes: []string{"linguist-vendored", "linguist-generated", "linguist-language", "gitlab-language"},
+		Attributes: []string{"linguist-vendored", "linguist-generated", "linguist-language", "gitlab-language", "linguist-documentation", "linguist-detectable"},
 		Repo:       repo,
 		IndexFile:  indexFilename,
 		WorkTree:   worktree,
@@ -316,3 +317,23 @@ func (repo *Repository) CheckAttributeReader(commitID string) (*CheckAttributeRe
 
 	return checker, deferable
 }
+
+// true if "set"/"true", false if "unset"/"false", none otherwise
+func attributeToBool(attr map[string]string, name string) optional.Option[bool] {
+	if value, has := attr[name]; has && value != "unspecified" {
+		switch value {
+		case "set", "true":
+			return optional.Some(true)
+		case "unset", "false":
+			return optional.Some(false)
+		}
+	}
+	return optional.None[bool]()
+}
+
+func attributeToString(attr map[string]string, name string) optional.Option[string] {
+	if value, has := attr[name]; has && value != "unspecified" {
+		return optional.Some(value)
+	}
+	return optional.None[string]()
+}
diff --git a/modules/git/repo_language_stats_gogit.go b/modules/git/repo_language_stats_gogit.go
@@ -11,6 +11,7 @@ import (
 	"strings"
 
 	"code.gitea.io/gitea/modules/analyze"
+	"code.gitea.io/gitea/modules/optional"
 
 	"github.com/go-enry/go-enry/v2"
 	"github.com/go-git/go-git/v5"
@@ -57,25 +58,47 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 			return nil
 		}
 
-		notVendored := false
-		notGenerated := false
+		isVendored := optional.None[bool]()
+		isGenerated := optional.None[bool]()
+		isDocumentation := optional.None[bool]()
+		isDetectable := optional.None[bool]()
 
 		if checker != nil {
 			attrs, err := checker.CheckPath(f.Name)
 			if err == nil {
-				if vendored, has := attrs["linguist-vendored"]; has {
-					if vendored == "set" || vendored == "true" {
-						return nil
-					}
-					notVendored = vendored == "false"
+				isVendored = attributeToBool(attrs, "linguist-vendored")
+				if isVendored.ValueOrDefault(false) {
+					return nil
+				}
+
+				isGenerated = attributeToBool(attrs, "linguist-generated")
+				if isGenerated.ValueOrDefault(false) {
+					return nil
 				}
-				if generated, has := attrs["linguist-generated"]; has {
-					if generated == "set" || generated == "true" {
-						return nil
+
+				isDocumentation = attributeToBool(attrs, "linguist-documentation")
+				if isDocumentation.ValueOrDefault(false) {
+					return nil
+				}
+
+				isDetectable = attributeToBool(attrs, "linguist-detectable")
+				if !isDetectable.ValueOrDefault(true) {
+					return nil
+				}
+
+				hasLanguage := attributeToString(attrs, "linguist-language")
+				if hasLanguage.Value() == "" {
+					hasLanguage = attributeToString(attrs, "gitlab-language")
+					if hasLanguage.Has() {
+						language := hasLanguage.Value()
+						if idx := strings.IndexByte(language, '?'); idx >= 0 {
+							hasLanguage = optional.Some(language[:idx])
+						}
 					}
-					notGenerated = generated == "false"
 				}
-				if language, has := attrs["linguist-language"]; has && language != "unspecified" && language != "" {
+				if hasLanguage.Value() != "" {
+					language := hasLanguage.Value()
+
 					// group languages, such as Pug -> HTML; SCSS -> CSS
 					group := enry.GetLanguageGroup(language)
 					if len(group) != 0 {
@@ -85,28 +108,14 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 					// this language will always be added to the size
 					sizes[language] += f.Size
 					return nil
-				} else if language, has := attrs["gitlab-language"]; has && language != "unspecified" && language != "" {
-					// strip off a ? if present
-					if idx := strings.IndexByte(language, '?'); idx >= 0 {
-						language = language[:idx]
-					}
-					if len(language) != 0 {
-						// group languages, such as Pug -> HTML; SCSS -> CSS
-						group := enry.GetLanguageGroup(language)
-						if len(group) != 0 {
-							language = group
-						}
-
-						// this language will always be added to the size
-						sizes[language] += f.Size
-						return nil
-					}
 				}
 			}
 		}
 
-		if (!notVendored && analyze.IsVendor(f.Name)) || enry.IsDotFile(f.Name) ||
-			enry.IsDocumentation(f.Name) || enry.IsConfiguration(f.Name) {
+		if (!isVendored.Has() && analyze.IsVendor(f.Name)) ||
+			enry.IsDotFile(f.Name) ||
+			(!isDocumentation.Has() && enry.IsDocumentation(f.Name)) ||
+			enry.IsConfiguration(f.Name) {
 			return nil
 		}
 
@@ -115,12 +124,10 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 		if f.Size <= bigFileSize {
 			content, _ = readFile(f, fileSizeLimit)
 		}
-		if !notGenerated && enry.IsGenerated(f.Name, content) {
+		if !isGenerated.Has() && enry.IsGenerated(f.Name, content) {
 			return nil
 		}
 
-		// TODO: Use .gitattributes file for linguist overrides
-
 		language := analyze.GetCodeLanguage(f.Name, content)
 		if language == enry.OtherLanguage || language == "" {
 			return nil
@@ -138,7 +145,7 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 			included = langtype == enry.Programming || langtype == enry.Markup
 			includedLanguage[language] = included
 		}
-		if included {
+		if included || isDetectable.ValueOrDefault(false) {
 			sizes[language] += f.Size
 		} else if len(sizes) == 0 && (firstExcludedLanguage == "" || firstExcludedLanguage == language) {
 			firstExcludedLanguage = language
diff --git a/modules/git/repo_language_stats_nogogit.go b/modules/git/repo_language_stats_nogogit.go
@@ -12,6 +12,7 @@ import (
 
 	"code.gitea.io/gitea/modules/analyze"
 	"code.gitea.io/gitea/modules/log"
+	"code.gitea.io/gitea/modules/optional"
 
 	"github.com/go-enry/go-enry/v2"
 )
@@ -88,25 +89,47 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 			continue
 		}
 
-		notVendored := false
-		notGenerated := false
+		isVendored := optional.None[bool]()
+		isGenerated := optional.None[bool]()
+		isDocumentation := optional.None[bool]()
+		isDetectable := optional.None[bool]()
 
 		if checker != nil {
 			attrs, err := checker.CheckPath(f.Name())
 			if err == nil {
-				if vendored, has := attrs["linguist-vendored"]; has {
-					if vendored == "set" || vendored == "true" {
-						continue
-					}
-					notVendored = vendored == "false"
+				isVendored = attributeToBool(attrs, "linguist-vendored")
+				if isVendored.ValueOrDefault(false) {
+					continue
+				}
+
+				isGenerated = attributeToBool(attrs, "linguist-generated")
+				if isGenerated.ValueOrDefault(false) {
+					continue
 				}
-				if generated, has := attrs["linguist-generated"]; has {
-					if generated == "set" || generated == "true" {
-						continue
+
+				isDocumentation = attributeToBool(attrs, "linguist-documentation")
+				if isDocumentation.ValueOrDefault(false) {
+					continue
+				}
+
+				isDetectable = attributeToBool(attrs, "linguist-detectable")
+				if !isDetectable.ValueOrDefault(true) {
+					continue
+				}
+
+				hasLanguage := attributeToString(attrs, "linguist-language")
+				if hasLanguage.Value() == "" {
+					hasLanguage = attributeToString(attrs, "gitlab-language")
+					if hasLanguage.Has() {
+						language := hasLanguage.Value()
+						if idx := strings.IndexByte(language, '?'); idx >= 0 {
+							hasLanguage = optional.Some(language[:idx])
+						}
 					}
-					notGenerated = generated == "false"
 				}
-				if language, has := attrs["linguist-language"]; has && language != "unspecified" && language != "" {
+				if hasLanguage.Value() != "" {
+					language := hasLanguage.Value()
+
 					// group languages, such as Pug -> HTML; SCSS -> CSS
 					group := enry.GetLanguageGroup(language)
 					if len(group) != 0 {
@@ -116,29 +139,14 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 					// this language will always be added to the size
 					sizes[language] += f.Size()
 					continue
-				} else if language, has := attrs["gitlab-language"]; has && language != "unspecified" && language != "" {
-					// strip off a ? if present
-					if idx := strings.IndexByte(language, '?'); idx >= 0 {
-						language = language[:idx]
-					}
-					if len(language) != 0 {
-						// group languages, such as Pug -> HTML; SCSS -> CSS
-						group := enry.GetLanguageGroup(language)
-						if len(group) != 0 {
-							language = group
-						}
-
-						// this language will always be added to the size
-						sizes[language] += f.Size()
-						continue
-					}
 				}
-
 			}
 		}
 
-		if (!notVendored && analyze.IsVendor(f.Name())) || enry.IsDotFile(f.Name()) ||
-			enry.IsDocumentation(f.Name()) || enry.IsConfiguration(f.Name()) {
+		if (!isVendored.Has() && analyze.IsVendor(f.Name())) ||
+			enry.IsDotFile(f.Name()) ||
+			(!isDocumentation.Has() && enry.IsDocumentation(f.Name())) ||
+			enry.IsConfiguration(f.Name()) {
 			continue
 		}
 
@@ -170,7 +178,7 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 				return nil, err
 			}
 		}
-		if !notGenerated && enry.IsGenerated(f.Name(), content) {
+		if !isGenerated.Has() && enry.IsGenerated(f.Name(), content) {
 			continue
 		}
 
@@ -193,13 +201,12 @@ func (repo *Repository) GetLanguageStats(commitID string) (map[string]int64, err
 			included = langType == enry.Programming || langType == enry.Markup
 			includedLanguage[language] = included
 		}
-		if included {
+		if included || isDetectable.ValueOrDefault(false) {
 			sizes[language] += f.Size()
 		} else if len(sizes) == 0 && (firstExcludedLanguage == "" || firstExcludedLanguage == language) {
 			firstExcludedLanguage = language
 			firstExcludedLanguageSize += f.Size()
 		}
-		continue
 	}
 
 	// If there are no included languages add the first excluded language
diff --git a/tests/integration/linguist_test.go b/tests/integration/linguist_test.go
